Exploring AI Through a Racial Lens with Marie Smith from Data360
Season 2, Episode 42, Mar 23, 2023, 11:18 AM
This week, Paul captains the podcast ship alone to talk to Marie Smith, Co-Founder and Chief Information Officer at Data360 on the intersection between AI, data, and racial equity.
Marie is a self-confessed “super nerd supreme” who started coding at just nine years old. She began her career as an engineer at AOL, then worked at Universal Studios and several tech firms before co-founding Data360 eight years ago.
The first article we discuss is from Forbes. It tackles the intersection between AI and Diversity, Equity, and Inclusion (DEI) and asks whether ChatGPT will make DEI communications from leaders inauthentic.
“It’s a matter of fairly sourcing data,” Marie tells us. She explains that sources that are widely regarded as authoritative tend to be biased, so there’s a need to use wider, more inclusive data sets to train AI.
The second article from IBM talks about the need to account for racial bias in machine learning models and the idea that data can carry implicit bias. We asked Marie what she thinks about how to remove racial bias from machine learning models.
“Who do you have on the team?” Marie asks, emphasizing the need for diversity among teams developing and training AI models. She then goes on to talk about how her background in anthropology influences how she thinks about AI, machine learning, and removing bias.
“You need four approaches: qualitative, quantitative, latitudinal, longitudinal,” Marie explains, “What we mostly have now is a quantitative approach, but there isn’t necessarily always a qualitative approach.”
The third piece from Black Girl Nerds asks whether AI can improve Black women’s lives. “AI apps and software have definitely improved the quality of my life,” Marie says. “Our company also works with Black women through our online incubator, and I would say yes, it does improve their lives.”
Marie shares some of the challenges she faces as a Black woman in a leadership role at a tech company before we move on to Data360’s work helping small business owners automate their business operations and increase their data literacy.
Finally, into what Marie thinks the evillest tech app is and her predictions for the future of tech.
To learn more about Data360: https://www.data360.solutions/home
The first article we discuss is from Forbes. It tackles the intersection between AI and Diversity, Equity, and Inclusion (DEI) and asks whether ChatGPT will make DEI communications from leaders inauthentic.
“It’s a matter of fairly sourcing data,” Marie tells us. She explains that sources that are widely regarded as authoritative tend to be biased, so there’s a need to use wider, more inclusive data sets to train AI.
The second article from IBM talks about the need to account for racial bias in machine learning models and the idea that data can carry implicit bias. We asked Marie what she thinks about how to remove racial bias from machine learning models.
“Who do you have on the team?” Marie asks, emphasizing the need for diversity among teams developing and training AI models. She then goes on to talk about how her background in anthropology influences how she thinks about AI, machine learning, and removing bias.
“You need four approaches: qualitative, quantitative, latitudinal, longitudinal,” Marie explains, “What we mostly have now is a quantitative approach, but there isn’t necessarily always a qualitative approach.”
The third piece from Black Girl Nerds asks whether AI can improve Black women’s lives. “AI apps and software have definitely improved the quality of my life,” Marie says. “Our company also works with Black women through our online incubator, and I would say yes, it does improve their lives.”
Marie shares some of the challenges she faces as a Black woman in a leadership role at a tech company before we move on to Data360’s work helping small business owners automate their business operations and increase their data literacy.
Finally, into what Marie thinks the evillest tech app is and her predictions for the future of tech.
To learn more about Data360: https://www.data360.solutions/home